Mc

Ea

b

c

a

ARRAA

KTPH

1

sbiutpictmtplmmtt

vT

0d

Journal of Chromatography A, 1217 (2010) 2138–2146

Contents lists available at ScienceDirect

Journal of Chromatography A

journa l homepage: www.e lsev ier .com/ locate /chroma

icro-bore titanium housed polymer monoliths for reversed-phase liquidhromatography of small molecules

katerina P. Nesterenkoa, Pavel N. Nesterenkob, Damian Connollya, Flavie Lacroixc, Brett Paull a,∗

Irish Separation Science Cluster, National Centre for Sensor Research, Dublin City University, Glasnevin, Dublin 9, IrelandAustralian Centre for Research on Separation Science, University of Tasmania, Hobart, Tasmania 7001, AustraliaCentre d’études superieures d’ingénieurs, 297, Rue de Vaugirard-75015, Paris, France

r t i c l e i n f o

rticle history:eceived 5 November 2009eceived in revised form 14 January 2010ccepted 2 February 2010vailable online 10 February 2010

a b s t r a c t

A new method for the fixation of polymethacrylate monoliths within titanium tubing of up to 0.8 mm I.D.for use as a chromatographic column under elevated temperatures and pressures is described. The prepa-ration of butyl methacrylate–ethylene dimethacrylate-based monolithic stationary phases with desiredporous structures was achieved within titanium tubing with pre-oxidised internal walls. The oxidisedtitanium surface was subsequently silanised with 3-trimethoxysilylpropyl methacrylate resulting in tight

eywords:itanium columnolymer monolithigh pressure reversed-phase HPLC

bonding of butyl methacrylate porous monolith to the internal walls, providing stationary phase stabilityat column temperatures up to 110 ◦C and at operating column pressure drops of >28 MPa. The titaniumhoused monoliths exhibited a uniform and dense porous structure, which provided peak efficienciesof up to 59,000 theoretical plates per meter when evaluated for the separation of small molecules inreversed-phase mode, under optimal conditions (achieved at 15 �L/min and temperature of 110 ◦C fornaphthalene with a retention factor, k = 0.58). The developed column was applied to the reversed-phase

ext m

isocratic separation of a t

. Introduction

Based upon their reported high permeability and acceptableeparation efficiencies, the use of monolithic porous media has noween demonstrated in most modes of chromatographic analysis

ncluding GC [1], HPLC [2], IC [3,4], CEC [5] and SPE [6]. Dependingpon the scale and specific goals of the chromatographic applica-ion monolithic porous bodies of various geometry and shape can berepared and housed, with the most common applications involv-

ng monolithic rods [7] and disks [8], and monoliths housed withinapillaries [9], micro-channels [10] and pipette tips [11]. However,here remains some significant size-related limitations in the use of

onolithic phases for chromatographic applications, which are stillhe subject of a considerable body of research. For example, organicolymers (polystyrene-divinylbenzene (PS-DVB), polymethacry-

ate (PMA), etc.) and various inorganic oxides represent the two

ain material classes which have been used in the preparation ofonolithic porous columns for liquid chromatography. However,

he complex combination of different physico-mechanical proper-ies and limitations of the above substrates, such as varying degrees

∗ Corresponding author at: National Centre for Sensor Research, Dublin City Uni-ersity, School of Chemical Sciences, Glasnevin, Dublin 9, Ireland.el.: +353 1 7005060; fax: +353 1 7005503.

E-mail address: Brett.Paull@dcu.ie (B. Paull).

021-9673/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.chroma.2010.02.007

ixture of pesticides.© 2010 Elsevier B.V. All rights reserved.

of rigidity and fragility (cracking, shrinking and swelling), pH andtemperature stability, surface area and pore geometry, and walladhesion chemistry, places some practical limits on the size andgeometry of columns which can be produced, especially at the levelof commercial production.

Rigid silica-based rods of diameter greater than 2 mm andlength 10–15 cm, can be prepared by established sol–gel tech-nology followed by calcination at high temperatures. Subsequentthermo-shrinking based cladding of these rods is used for thepreparation of commercial HPLC columns [12]. On the capillary endof the scale, the significant shrinking of hydrated silica gel duringcalcination, and the fragility of the final silica monolith produced,limits the possibilities of in situ preparation of columns in standardfused silica (FS) capillaries of I.D. less than 100 �m. However, theactivated/hydroxylated internal walls of FS capillaries are involvedin the sol–gel process of the monolith body formation, providingstrong bonding of the monolith body within such FS capillaries ofI.D. of >100 �m.

In situ polymerisation in various sizes of FS capillaries is a com-mon approach for the preparation of organo-polymer monolithiccolumns [2,13]. However, difficulties again arise in the prepara-

tion of long columns of internal diameter greater than 1–2 mm.The stable bonding of the polymer rod within the column housingis a key challenge, particularly for higher pressure and temper-ature modes of chromatography. Current commercially availableorgano-polymer-based columns can generally only operate at

matog

bTmoismstumi

dasccstctsvocgorga

tctdcpmibihvsadtbo

pttismeWfcsbofs

E.P. Nesterenko et al. / J. Chro

ackpressures up to 20 MPa and temperatures up to 70 ◦C [14].his of course depends upon the rigidity of the organo-polymeratrix, the internal wall surface area and, not less importantly,

n the strength of bonds between the polymer matrix and thenternal walls of the column housing. Thus, Courtois et al. [15]tudied various techniques for surface pretreatment and attach-ent of anchoring groups to fused silica capillaries surface and

howed that silanisation technique is of extreme importance ashe strength of monolith attachment depends on the chemistrytilised to obtain fixation. Other than that, the strength of attach-ent also depends on the chemical nature of the column housing

tself.An advantage of FS capillaries as a column housing is the high

ensity of silanol groups on the capillary wall which can be fullyctivated after alkali or acidic treatment [16]. The fully activatedilica surface contains 4.6 silanol groups/nm2. Obviously, in thease of the in situ preparation of a silica monolith within a FSapillary, all available silanols are involved in the sol–gel synthe-is, which provides very strong bonding of the silica structure tohe wall. Similar bonding of organo-polymer monoliths within FSapillaries needs activation of the internal walls with silanes con-aining reactive groups. Due to steric hindrance only 50–60% ofurface silanols can be converted into reactive functional groupsia such treatment although this amount is also sufficient to holdrgano-polymer monoliths within the FS capillary under standardhromatographic conditions. The type of the reactive functionalroup used for the bonding of organo-polymer monoliths dependsn the type of polymer monolith being prepared. Epoxy- or oxi-ane groups are often used for the preparation and bonding oflycidylmethacrylate (GMA) based polymers in FS capillaries [17],nd vinyl- groups utilised for PS-DVB polymers.

Clearly, the covalent fixation of the organo-polymer monolitho the silica walls is an extremely useful advantage of using a FSapillary column format in monolithic chromatographic applica-ions. However, the use of glass column housings of larger internaliameters has not been significantly developed due to problems ofolumn handling and strength issues related to operation at higherressures. Additionally, poly-ether-ether ketone (PEEK) and com-on stainless steel (SS) tubes have also been largely avoided, as

n both cases the arrangement of chemical bridges between theulk monolith and walls is not trivial. In the case of SS the mod-

fication of the surface with silanes is not viable due to the poorydrolytic stability of the Me–O–Si–C bonding, where Me includesarious metals such as nickel, iron, chromium from the stainlessteel composition. The introduction of reactive functional groupst the surface of PEEK is also highly problematic and is not yet wellemonstrated in the literature. The applicability of the aforemen-ioned cladding technique used for silica-based rods is restrictedy the relative softness of the organo-polymer monoliths and cannly be used in the production of disks or relatively short columns.

In this paper a new alternative column housing for organo-olymer monoliths is presented, which can help overcome some ofhe above limitations. For the first time surface oxidised titaniumubing of 0.8 mm I.D. has been evaluated as a suitable column hous-ng material for organo-polymer monoliths, demonstrating highlytable surface to monolith bonding strength, structural and ther-al stability of the in-column polymerised monoliths, and very

ncouraging chromatographic performance for small molecules.ithin the titanium tubing a thin layer of titanium oxide can be

ormed by the simple oxidation of the titanium wall surface, whichan be further modified using silane chemistry similarly to a FS

urface, providing a stable Ti–O–Si-type of bonding, which cane both stable and readily amenable to further bonding with thergano-polymer matrix. The in-column polymerisation parametersor formation of monoliths of stable pore structure for efficienteparation of small molecules are also presented.

r. A 1217 (2010) 2138–2146 2139

2. Experimental

2.1. Chemicals and reagents

Butyl methacrylate, ethylene dimethacrylate, 1-decanol and3-trimethoxysilylpropyl methacrylate were all purchased fromSigma–Aldrich (Gillingham, UK), as were sodium hydroxide andhydrochloric acid. The thermal initiator 1,1′-azobisizobutyronitrile(AIBN) was obtained from DuPont (Le Grand Sacconex, Switzer-land). Solvents used for the synthesis, washing of the preparedmonolith and for the preparation of eluents, namely acetonitrile(ACN), methanol, acetone and toluene were purchased from LabScan (Gliwice, Poland). Titanium tubes (100 mm × 0.8 mm I.D.)were obtained from Carl Stuart Ltd. (Dublin, Ireland). Deionisedwater purified by a Milli-Q system (Millipore, Bedford, USA)was utilised throughout the experiments. The standard solu-tions, comprising of uracil, acetophenone, benzene, toluene andnaphthalene were obtained from Phenomenex (Torrance, CA,USA).

2.2. Instrumentation

A EuroTherm 91e furnace (Carbolite Furnaces, Hope, UK) wasused for the thermal oxidation of the titanium tube surface inorder to provide a sufficient surface layer of titanium oxide. ASputterCoater S150B (BOC Edwards, Sussex, UK) was utilised forcoating the subsequent titanium housed polymer monolith sam-ple prior to scanning electron microscopy analysis, which wasperformed on a S-3400N instrument (Hitachi, Maidenhead, UK).For the chromatographic study, an Ultimate 3000 Rapid Separa-tion HPLC system was used, comprising of a HPG-3x00RS binarypump, WPS-3000RS autosampler, TCC-3000 RS thermostated col-umn compartment (which allows the use of temperatures of up to110 ◦C) and a MWD-3000RS multiple wavelength detector (Dionex,Sunnyvale, CA, USA).

2.3. Preparation of titanium housed monolithic columns

Titanium tubes (100 mm × 0.8 mm I.D.) were placed in the fur-nace and then heated from room temperature to 500 ◦C, and keptat this temperature for 6 h. The furnace was then switched offand allowed to cool down slowly to room temperature. This wasfollowed by activation of hydroxide groups on the surface of theoxidised titanium tube to provide reactive groups for the subse-quent silanisation. This was achieved through washing with 0.2 MNaOH for 1 h. The tubes were then washed with deionised waterfor 30 min and a further washing step with 0.2 M HCl for 2 h. Fol-lowing this the tubes were again washed with deionised waterfor 15 min followed by acetone and then left to dry for 1 h at90 ◦C.

After the above oxidation and activation steps, the walls ofthe column were vinylised in order to provide covalent attach-ment points for the polymer monolith to the titanium surface.This was performed by reacting the column walls with 3-trimethoxysilylpropyl methacrylate, which acts as a linker betweenthe inorganic titanium surface and organic polymer material.The pre-treated titanium tubes were silanised with 50 wt% 3-trimethoxysilylpropyl methacrylate in toluene at 90 ◦C for 24 h. Nopolymerisation inhibitors were added.

Finally, the butyl methacrylate-based monolithic stationaryphase was produced using an adapted procedure of Geiser,

previously shown to be both robust and reproducible [18].Methacrylate-based monolithic stationary phases were chosenfor the study due to their rigid properties and ease of prepara-tion through direct co-polymerisation of the functional monomerand cross-linker. The monomer mixture consisted of 24 wt% of

2 matog

bdwddotspTtafiNtmswp(tipasptTctaamtd

3

3

oaambTtlwtcsiaPow(ar

saa

140 E.P. Nesterenko et al. / J. Chro

utyl methacrylate, 16 wt% ethylene dimethacrylate, 60 wt% 1-ecanol and 1 wt% AIBN (in respect to monomers). The initiatoras weighed into the mixture vessel first, then the porogen (1-ecanol) was added and the mixture was sonicated for 60 min toissolve AIBN (AIBN was first dissolved in the porogen, due to thebserved limited solubility of AIBN in the polymerisation mixture,hus avoiding any possible negative effects from the presence ofolid phase in the polymerisation mixture, and also avoiding anyossible starting polymerisation due to heating during sonication).hen the rest of the reagents were added and the mixture was vor-exed for 5 min. After that the mixture was centrifuged for 13 mint 10,000 r/min. and 80% of the supernatant was taken and used toll an oxidised and silanised titanium tube, which was placed inMR test tube (no sediment was observed, however, centrifuga-

ion was carried out in order to remove any possible solid phaseicro-contaminants, in order to avoid their affect on the polymeri-

ation process). Prior filling the mould, polymerisation mixtureas deaerated for 10 min under the flow of nitrogen. The pre-ared tube was placed in the water-bath at room temperature20 ◦C), the temperature was slowly increased using a tempera-ure gradient from 20 to 50 ◦C over a period of up to 1 h, in smallncrements of ∼5 ◦C every 10 min and from 50 to 55 ◦C over aeriod of 50 min in small increments of ∼1 ◦C. Once the temper-ture reached 55 ◦C, it was kept constant for 8 h. After the end ofynthesis, the prepared column was left to cool down at room tem-erature overnight. To remove the column from the glass NMR testube, it was carefully crushed followed by removal of the glass.he bulk monolith outside the column was then cut off and theolumn cleaned with fine sand paper. After that appropriate fit-ings were attached to the column and it was washed with ACNt a flow rate of 1 �L/min for approximately 100 min and thent 10 �L/min for at least 60 min. During the removal of residualonomer mixture and porogen from the pores of the monolith,

he backpressure initially increased to 20 MPa and then droppedown to 3.9 MPa.

. Results and discussion

.1. Oxidation of titanium column housings

As titanium is highly reactive and has a very high affinity forxygen, titanium oxide films on the surface form spontaneouslynd instantly when metal is exposed to air. The nature, compositionnd thickness of the oxide layer on the surface depends on environ-ental conditions. In most environments the oxide is typically TiO2,

ut can also consist of a mixture of oxides, including TiO2, Ti2O3 andiO [19]. These naturally formed films on the surface of titanium areypically less than 10 nm thick, so in order to provide a sufficientayer of titanium oxide on the inner surface of the titanium tube, it

as oxidised prior to silanisation. Titanium surfaces can be oxidisedhrough chemical oxidation or thermal oxidation. Though chemi-al oxidation provides a consistent and reproducible titanium oxideurface layer, thermal oxidation was chosen for this procedure ast does not require the use of such chemicals as sulphuric acidnd hydrogen peroxide [20]. Additionally, studies performed byadma et al. [21] showed that at temperatures below 900 ◦C andxidation duration over 3 h, a layer of titanium oxide was formed,hich consisted of 98–100% TiO2, depending on depth of the layer

0–1 �m, with the percentage of TiO2 decreasing with layer depth),nd that the composition of the oxide in this surface layer was both

eproducible and consistent.

The temperature of the titanium oxidation procedure was veryignificant. At oxidation temperatures below 600 ◦C a chemicallyctive modification of titania (anatase) exists, while at temper-tures over 620–650 ◦C, a chemically inert modification–rutile is

r. A 1217 (2010) 2138–2146

formed. As the oxidation step is followed by activation of the OH–groups (typically the surface concentration of Ti–OH groups wouldbe 4.3–4.8 groups/nm2 [19]) and attachment of vinyl functionalgroups, it is important to have an active form of TiO2 on the surface.So the thermal oxidation was performed at 500 ◦C for 6 h.

3.2. Optimisation of polymerisation temperature and duration

Another crucial step in the formation of organo-polymer mono-liths with uniform porosity in titanium housing, is optimisation ofboth the temperature and duration of polymerisation. Initially thepolymerisation step was performed at 60 ◦C for 24 h based uponprevious studies using FS capillaries [22]. However, when the col-umn was connected to the HPLC system, it was not possible to pumpthrough it, with backpressures >20 MPa at 1 �L/min. No reductionin backpressure was observed after 72 h of constant washing withACN. SEM imaging confirmed that the porosity of the stationaryphase was very low. A reason for the formation of a monolith ofsuch high density could be the high thermal conductivity of tita-nium, which is 21.9 W/m/K [23] (for FS this is 1.3 W/m/K [24], andfor PEEK this value is only 0.25 W/m/K [25]). This high thermal con-ductivity results in faster heat dissipation in the titanium column.As a consequence the monomer mixture would heat up faster, pro-viding a more rapid polymerisation and a higher temperature insidethe column [26].

It is well known that in order to obtain fine control of mono-lith porosity, reaction parameters such as nature of porogen, theratio of monomer to cross-linker in the mixture, polymerisationtemperature and time can be varied. Here, the aim of the workwas successful attachment of the polymer monolith to the titaniumhousing, together with demonstration of the resultant monolithiccolumn chromatographically under elevated column temperatureand pressure conditions. Therefore, in this study only the polymeri-sation temperature and duration were studied in detail.

The reaction rate for free-radical polymerisation is not a sim-ple function of temperature, as the process consists of several steps[2,27,28]. The most important is the initiation rate, which is highlydependant on temperature, since the half-life of initiators decreaseswith increases in temperature. As a result, the decomposition rateof the initiator, the number of radicals produced, the formation ofmore globules, and consequently the overall polymerisation rateare higher at elevated temperatures. As the formation of new poly-merisation centres is faster than the growth of globules, the supplyof monomers runs low fast and the number of globules is large,but their size stays small, leading to smaller voids between glob-ules. So essentially, higher polymerisation temperatures result insmaller pore sizes.

The duration of polymerisation also has an affect upon thepore size of the resultant monolith. With longer polymerisationtimes, monolith globules increase in size resulting in smaller voidsbetween them, and as a result leading to smaller pores and higherdensity of the monolith [2].

For initial investigations into the temperature and polymeri-sation duration effects on the porosity of the monoliths, butylmethacrylate–ethylene dimethacrylate monoliths were preparedin bulk (i.e. not in titanium column housing), pre-treated asdescribed in the experimental section. The polymerisation time wasvaried from 7 to 24 h at a constant temperature of 60 ◦C. In addi-tion, the polymerisation temperature was varied from 54 to 60 ◦Cfor a fixed time of 7 h. SEM images of the prepared monoliths were

obtained and sizes of pores and polymer globules were estimated.The semi-quantitative results are presented in Table 1. As it can beseen, the results obtained correspond with data presented in pre-vious studies [2,7,29], i.e. higher polymerisation temperatures andlonger reaction times lead to smaller pore sizes.

E.P. Nesterenko et al. / J. Chromatogr. A 1217 (2010) 2138–2146 2141

Table 1The influence of polymerisation temperature and duration on approximate flow-through pore size.

Polymerisationtemperature, ◦C atfixed time (7 h)

dpore, �ma Polymerisation time,h at fixedtemperature (60 ◦C)

dpore, �ma

54 1.18 ± 0.59 7 0.78 ± 0.2455 0.78 ± 0.31 9 0.43 ± 0.2356 0.63 ± 0.37 12 0.34 ± 0.2057 0.71 ± 0.57 24 0.27 ± 0.16

3p

stbchegcoclpdc[awtcit

motgv

Fmme

Fig. 2. SEM images of titanium housed monoliths formed using thermal gradient

58 0.54 ± 0.2560 0.43 ± 0.3

a Calculated for n = 30 pores.

.3. Effect of gradient temperature programming duringolymerisation

The preparation of monoliths with a homogeneous poroustructure in capillaries is readily achieved, whereas scaling uphe diameter of the monolith while still maintaining good wall-onding is somewhat more challenging. The initial titanium housedolumns prepared at the early stages of this work were found toave a non-uniform porosity distribution when cross-sections werexamined by SEM, as shown in clearly in Fig. 1. The radial homo-eneity of porosity is clearly quite poor, with larger pores in theentral core of the monolith (section A) and smaller pores in theuter shell formed near the column walls (section B). These resultsorrespond with the data obtained in [26] for the preparation ofarger size polymer monoliths. It is likely here that the non-uniformorosity resulted from the uncontrolled increase in temperatureuring the preparation of large-diameter monoliths, as a signifi-ant exotherm can be observed during the polymerisation process26] (depending on the type of the free-radical thermal initiatornd the dimensions of the column, the exotherm can be up to 60 ◦C,hich obviously has an effect on the polymerisation kinetics and

he resultant structure of the monolith). However, an alternativeause of such non-homogeneity could also involve some degree ofnternal surface catalysed decomposition of AIBN, thus acceleratinghe initiation rate close to the column surface.

To investigate this further here, it was decided to run the poly-erisation at a rate slow enough to allow dissipation of the heat

f polymerisation, in order to attempt to provide the formation ofhe monolith with uniform porosity. Therefore, a slow temperatureradient was used during preparation of the monolith, which pro-ided uniform heating of the column housing and the monomer

ig. 1. Initial SEM image of radially inhomogeneous structure of a monolith ther-ally polymerised in 0.8 mm I.D. titanium tubing housing. Zone A corresponds to aore porous central portion of the monolith and Zone B corresponds to a non-porous

xterior of the monolith.

polymerisation. (a) SEM image showing polymer monolith bound to the inner wallof titanium column and Zone A corresponds to the polymer monolith, Zone B isthe titanium column housing. (b) SEM image showing dense highly homogenousmonolithic structure.

mixture within (the application of a temperature gradient shouldnot change the structure of the polymer inside the titanium tube ifsuch a structure was related to surface catalysis effects rather thantemperature, unless the catalysis kinetics were also affected). Thetemperature gradient applied was selected in such way in order toprovide gradual heating. As polymerisation would not start at tem-peratures under 50 ◦C, up to that temperature the heating rate wasrelatively fast, and to provide better equilibration at temperaturesabove 50 ◦C, the heating rate was then significantly reduced. Theimage of the monolith, prepared under these conditions is shownin Fig. 2(b). The absence of a significant exotherm during slow poly-merisation [26] correlates with the formation of the completelyhomogeneous porous structure obtained.

3.4. Scanning electron microscopy images of the monolith housedin titanium column

Using the optimised polymerisation protocols described above,a 10 mm × 0.8 mm I.D. monolith in titanium housing was preparedand imaged using SEM. Once the column was prepared, it was son-

icated in acetone for 60 min and then dried for 3 h at 60 ◦C. The drysample was sputtered with gold, forming a layer of approximately60 nm. The images obtained are shown in Fig. 2. Excellent com-prehensive attachment of the polymer monolith to the titanium

2 matog

cim

ufhe

b

F

wctef

F

wmtrf

chophaa

vmfltp(Cgt∼tmntt

3

tistfciheH

142 E.P. Nesterenko et al. / J. Chro

olumn wall is clearly seen in Fig. 2(a) and (b) permits an approx-mate evaluation of the porosity of the bulk dense monolith in the

iddle of the column.The bond strength of the polymer monolith to the titanium col-

mn housing is of crucial importance in this work and an importantactor in its suitability for applications under conditions generatingigh column pressure drops. Therefore, it is worth briefly consid-ring here the parameters involved.

Obviously, the force of interaction, Fint, between the monolithody and the internal walls is proportional to the interface area:

int = aS = 2a�rl,

here l and r are column length and radius and a is the adhesionoefficient. In the case of an ideal incompressible porous monolith,he force pushing the monolith out from the column, FP, can bexpressed through the pressure applied to the top of the column asollows:

P = �p�r2(1 − ε),

here �p is the column pressure drop and ε is the porosity of theonolith expressed as the ratio of the volume of pores in 1 g of

he monolith to its total volume. The monolith body would only beetained in the column housing (assuming no retaining supportingrit) when Fint > FP or at:

Fint

FP= 2a�rl

�p�r2(1 − ε)= 2al

�pr(1 − ε)≥ 1.

The above relationship states that under similar mobile phaseonditions, higher column stability or greater compatibility withigher operating pressures, can be obtained with longer columnsf smaller internal diameter. The other important parameters areorosity of the column and the adhesion coefficient, a. Clearly, theighest value of the coefficient a should result from comprehensivend homogeneous chemical bonding between the monolithic rodnd internal walls of the column housing.

The SEM images show that the density of the monolith wasery high, providing a relatively high surface area, though the per-eability of the column was expected to be relatively low. The

ow-through pore diameter was estimated from the SEM datao be approximately 0.54 ± 0.17 �m (n = 30), and the diameter ofolymer globules was found to be approximately 0.23 ± 0.08 �mn = 30). These results were compared with the data reported byauson et al. [35] and Vlakh and Tennikova [13]. The approximatelobules size of the polymer monoliths, calculated by authors ofhe present work from published SEM images and was found to be2.2 ± 0.6 �m and ∼2.0 ± 0.6 �m, respectively, which shows that

he organo-polymer monolith, obtained in the present work has auch more dense structure. The total pore volume for the tita-

ium housed column was estimated from the determined deadime value and was calculated to be 61% of the total volume ofhe column.

.5. Chromatographic evaluation of the titanium house monolith

Rapid and high throughput separations can be achieved eitherhrough using monolithic stationary phases with high permeabil-ty, allowing the use of elevated flow rates [30], or through usingub–2 �m particles [31], but the main disadvantage in the lat-er case is a very high backpressure. Another option to provideaster separations is the use of higher temperatures, as the vis-

osity of the eluent would reduce resulting in an overall decreasen backpressure, providing not only the possibility of the use ofigher flow rates, but also higher diffusion of the analyte in theluent and stationary phase, enabling improved mass transfer.owever, the use of aqueous based mobile phases at elevated

r. A 1217 (2010) 2138–2146

temperatures can affect the stability of silica-based monoliths,or the stability of the bond between polymer monoliths andthe walls of FS capillaries. A number of studies have been pub-lished which examine the stability of different stationary phasesat higher temperatures [32,33]. Andersen et al. [33] showed that apoly(styrene-divinybenzene) monolithic stationary phase was sta-ble over time (400 h) at 80 ◦C and no change in retention of analyteswas observed, while at 120 ◦C, a 20% decrease in retention factorwas observed after 100 h. The authors also reported that a smallvoid (<0.5 mm) was observed at the column inlet when the columnstested at 120 ◦C were opened after use. Eeltink et al. [34] alsodemonstrated similar results, where he demonstrated the appli-cation of poly(styrene-divinybenzene) monolith at 80 ◦C. Causonet al. [35] studied the stability of a poly(divinylbenzene) mono-lithic column when using pure water at temperatures up to 250 ◦Cand showed that the monolith was stable for up to 30 h of operat-ing at 220 ◦C. As for commercially available monolithic columns,their thermal stability range varies from 40 to 80 ◦C, for exam-ple, for the poly(styrene-divinylbenzene) based ProSwift columnsmaximum operating temperatures of 80 ◦C are recommended [36],while CIM disks, depending on the type of monolith housed, can beoperated at temperatures only up to 40–50 ◦C [37]. In the currentwork the titanium housed monolithic columns were constantly runat column temperatures of between 100 and 110 ◦C. During thischromatographic evaluation work, no significant change in reten-tion times for test analytes was observed, this amounting to ∼500 hof constant use.

The reversed-phase monolithic columns, prepared as describedabove were characterised using with a number of approaches, suchas column stability tests at high temperatures and backpressures,sample loading capacity, and backpressure dependence on the elu-ent flow rate.

3.5.1. Temperature, flow rate and column generatedbackpressures

Using ACN as the mobile phase, with a flow rate of 10 �L/min,column temperatures were slowly increased to 110 ◦C with incre-mental steps of 5 ◦C. At each step the column was left to equilibrateuntil constant backpressure was reached. A decrease in columngenerated backpressure was observed with an increase of the col-umn temperature, due to the decrease in mobile phase viscosity[38]. As the monoliths obtained in this work were very dense, thebackpressures were relatively high even at low flow rates. Thusincreasing column temperature was used to allow increased flowrates, to obtain optimum separation efficiency and decrease sepa-ration time. With an increase of the column temperature from 25to 110 ◦C, column generated backpressure dropped by 3.5 MPa.

To study the influence of the flow rate on the column backpres-sure, ACN and water were used as mobile phases, with a columntemperature of 100 ◦C. The flow rate was gradually increased withan increment of 1 �L/min and the backpressure was allowed tostabilise at each flow rate value. Fig. 3 shows that backpressurelinearly increases with the increase in flow rate, clearly demon-strating monolith and bonding stability over this range. It wasshown that the backpressure increases from approximately 1.0and 1.2 MPa for water and ACN at 1 �L/min, respectively, to 18.1and 24.9 MPa for the same mobile phases respectively at the flowrate of 22 �L/min. The higher pressure seen with ACN comparedto water is possibly related to a different degree of swelling ofthe prepared polymer monolith in the more non-polar solvent,which appears more significant than simply the relative viscosi-

ties of the two mobile phases. The same linear dependence ofbackpressure on the mobile phase flow rate was observed for abutyl methacrylate–ethylene dimethacrylate monolithic column(210 mm × 0.32 mm I.D.) with a similar composition (60% porogen,40% monomer mixture), and it was also shown that the backpres-

E.P. Nesterenko et al. / J. Chromatogr. A 1217 (2010) 2138–2146 2143

Fm

sa

3

rsaaaslfessewTod1f

uslbns

3

ecuis

Fig. 4. Overlaid 11 chromatograms (each 5th chromatogram) of 55 consecutiveinjections of a standard mixture on a butyl methacrylate–ethylene dimethacry-

ig. 3. Column generated backpressure dependence on flow rate at 100 ◦C withobile phases of ACN (R2 = 0.999) and water (R2 = 0.999).

ure increased from approximately 1.2 MPa at 1 �L/min to 3.0 MPat 4 �L/min using acetonitrile–water mixture of 65–35% [17].

.5.2. Separation reproducibilityIn order to investigate column stability and chromatographic

eproducibility under conditions of high temperature and pres-ure, a standard mixture of uracil, acetophenone, benzene, toluenend naphthalene was injected repeatedly and continuously over55-h period. Though the separation time for the mixture was

pproximately 10 min, the mixture was injected once per hour,o that a total of 55 repeat injections were carried out, equiva-ent to over 1200 column volumes. The resultant chromatogramsrom this reproducibility study, are presented in Fig. 4, showingvery 5th chromatogram from up to run 55, overlaid. As it can beeen, the column was stable over this period of time. The relativetandard deviation values (RSD) for such chromatographic param-ters as retention time (tr), peak asymmetry (As) and efficiency (N)ere found to be 0.8–1.1%, 1.1–3.4% and 1.4–5.6%, respectively.

he calculated RSD values were found to be smaller than thosebtained by Holdsvendova et al. [39] for methacrylate–ethyleneimethacrylate-based monolithic capillary columns, which were.7–4.8% for the retention factor (k), 2.7–3.9% for As and 1.6–16%or N.

Low RSD values indicate that after prolonged use of the col-mn at higher temperature and pressure, no changes in the columntructure occurred (e.g. cracks, voids) and that the polymer mono-ith did not detach from the inner column wall and was still stronglyound to it. It is also important to note that during this experimento change in the column backpressure was observed, which alsouggests that no detachment of the monolith occurred.

.5.3. Evaluation of separation efficiencyThe efficiency of the titanium housed monolithic column, was

valuated under varying eluent flow rates. A van-Deemter plot wasonstructed for naphthalene peak (100 ppm, 0.2 �L injection vol-me), with eluent flow rates ranging from 1 to 18 �L/min. The plot

s shown as Fig. 5. The velocity range was limited by the backpres-ure, although the data indicated the minimum H values were to be

late polymer monolith housed within a 100 mm × 0.8 mm I.D. titanium column.Mobile phase = 60% ACN–40% water, F = 10 �L/min, column temperature = 100 ◦C. UVdetection at 254 nm.

found at mobile phase velocities of between 0.6 and 0.8 mm/s (15and 18 �L/min). Under these flow conditions backpressures wererelatively high (e.g. 34 MPa at 18 �L/min). At the optimal flow rateof 15 �L/min, the efficiency calculated for the naphthalene peakwas ∼59,300 N/m.

The observed efficiency data over the range of mobile phasevelocities studies, compared favourably to that reported by Siouffi[40], for various polymer-based monoliths, and to that reported byMoravcova et al. [22], who studied van-Deemter’s curves for butylmethacrylate–ethylene dimethacrylate-based monolithic capillarycolumns prepared using different porogens.

The influence of the injection volume on peak efficiency wasalso studied. Chromatographic conditions were as follows: the elu-ent used was 90% acetonitrile–10% water, flow rate was 10 �L/minand the column temperature was 60 ◦C. For this study the 100 ppmnaphthalene solution was injected onto the column with injec-tion volumes ranging from 0.03 �L to 2 �L. Reduction of injectionvolume can reduce peak broadening since column overloading iseliminated, and as anticipated, Fig. 6 shows the increase in peakefficiency with a decrease in injection volume. Some small varia-tion in the trend was seen at the lowest injection volumes, possiblydue to uncertainty introduced from operation of the autosampler atthe very lowest of its range, although this was not deemed signifi-cant. Indirectly these results could be used to ascertain the loadingcapacity of the column used, which if taken as the injection vol-

ume resulting in a 10% reduction in efficiency from the maximumefficiency obtained, would be equal to a maximum recommendedinjection volume of 0.6 �L.

2144 E.P. Nesterenko et al. / J. Chromatogr. A 1217 (2010) 2138–2146

Table 2Dimensions and reported isocratic peak efficiencies of some organo-polymer monolithic columns in acetonitrile–water mobile phases at elevated temperatures and pressures.

Column L, cm I.D., mm Flow rate/cross-section,cm/min

Tmax, ◦C P, bar N theor. plates, m Ref.

MS/BVPE poly(p-methylstyrene–1,2-bis(p-vinylphenyl)ethane)

8 0.2 32 50 80 (25 ◦C) 25,000–35,000 [41]

BVPE poly(1,2-bis(p-vinylphenyl)ethane) 20 0.2 32 25 n/a 49,000–70,000 [42]ODMA/EDMA poly(n-octadecyl

methacrylate–ethylene dimethacrylate)100 0.25 6 80 15 20,000 [43]

Poly DVB poly(divinylbenzene) 23/42 0.25 20 200 n/a 6000a [35]

3

aeopouFttpoFcant

(

Fdwifi(

BuMA/EDMA poly(butylmethacrylate–ethylene dimethacrylate)

10 0.8 2

a Calculated from chromatogram from Ref. [33] for a 23 cm column.

.5.4. Separation of small moleculesFinally, a standard mixture of small organic molecules was sep-

rated on a prepared titanium housed monolithic column using anluent of 60% acetonitrile–40% water and a column temperaturef 110 ◦C. The obtained chromatogram is shown as Fig. 7(a). Therepared column was also used for the separation of the mixturef pesticides (Fig. 7(b)), but for comparative analysis of our col-mn with previously published work (below), data was taken fromig. 7(a). Clearly, the use of an applied gradient would improvehe separation of small molecules, particularly the separation ofhe mix of nine pesticides shown. However, here the separation ofesticides was included simply to show the potential applicationf the new column to small organic molecules in RP-HPLC mode.or polymeric monolithic phases, this represents a real on-goinghallenge, and so although the separation could be improved withgradient, it actually demonstrates more of the potential of the

ew monolithic phase by using just showing this isocratic separa-ion.

The following peak efficiencies were achieved for uracil∼59,000 N/m), acetophenone (∼52,000 N/m), benzene

ig. 5. van-Deemter plot for a titanium housed butyl methacrylate–ethyleneimethacrylate monolith, 100 mm × 0.8 mm I.D., mobile phase = 90% ACN–10%ater. Column temperature = 100 ◦C. Sample: naphthalene (100 ppm, 0.2 �L

njection volume). Data fitted using function y = a + bx + c/x giving calculated coef-cients a = 4.25 (±1.54) × 10−3 mm, b = 1.63 (±0.42) × 10−3 mm2/s and c = 2.82±0.08) × 10−2 ms.

110 196 59,000 Present work

(∼52,000 N/m), toluene (∼50,000 N/m) and naphthalene(∼49,000 N/m). These results were compared to those obtainedby Moravcova et al. [22] and Holdsvendova et al. [39], whoutilised butyl methacrylate–co–ethylene dimethacrylate-basedmonolithic capillary columns and observed peak efficienciesup to ∼34,000 N/m for benzene and ∼29,000 N/m for toluene[22] and ∼35,000 N/m for uracil, ∼32,000 N/m for benzene and∼28,000 N/m for toluene [39]. Trojer et al. [41] tested MS/BVPE(p-methylstyrene/1,2-bis(p-vinylphenyl)ethane) monolithic capil-lary columns for the separation of small molecules under gradientconditions and elevated temperature, as well as under isocraticconditions and ambient temperature. For the latter conditions,efficiencies of up to 35,000 N/m were achieved. Greiderer et al.[42] reported efficiencies of up to 70,000 N/m for small moleculeson a BVPE (poly(1,2-bis(p-vinylphenyl)ethane)) capillary columnin isocratic conditions at ambient temperature (Table 2). Ueki et al.[43] studied the behavior of methacrylate-base monoliths, namelyco-polymers of methacrylate esters with alkyl chains (C2–C18)and ethylene glycol dimethacrylate (EDMA). The best resultswere shown for the n-octadecyl methacrylate/EDMA monolithic

stationary phases, where the efficiency of up to 20,000 N/m wasachieved for the separation of 6 alkylbenzenes with elevatedcolumn temperatures (Table 2).

Therefore, in the work presented here with new titaniumhoused monoliths, the separation efficiencies are at least compa-

Fig. 6. The dependence of peak efficiency on the injection volume. Column: titaniumhoused butyl methacrylate–ethylene dimethacrylate monolith, 100 mm × 0.8 mmI.D., mobile phase = 60% ACN–40% water. Analyte – naphthalene, concentration –0.1 g/L.

E.P. Nesterenko et al. / J. Chromatog

Fig. 7. Separation of a (a) test mixture on a titanium housed butylmethacrylate–ethylene dimethacrylate polymer monolithic stationary phase.Column dimensions: 100 mm × 0.8 mm I.D. Mobile phase = 60% ACN–40% water,F = 10 �L/min, column temperature = 110 ◦C, column backpressure = 19.6 MPa and(b) the separation of pesticides: 1 – Paraquat, 2 – Aldicarb, 3 – 2-hydroxy-4-methoxy-benzene, 4 – 5-chlorosalicylaldehyde, 5 – naphthalene, 6 – Chlorpyrifos, 7– hexamethylbenzene, 8 – Dieldrin, 9 – DDT. Column: butyl methacrylate–ethylenedA2

re

4

t

[

[[

[[[

[[

[

[

[

[[[[[[[[[[[

[

[

imethacrylate polymer monolith, 100 mm × 0.8 mm I.D. Mobile phase = 60%CN–40% water, F = 10 �L/min, column temperature = 70 ◦C. UV detection at54 nm.

able to previously published work, and in some cases significantlyxceed previously published performance.

. Conclusions

The results presented herein demonstrate a novel approacho the preparation of larger scale polymer monolithic station-

[

[[[

r. A 1217 (2010) 2138–2146 2145

ary phases covalently bound to the titanium housing. It wasshown that columns prepared according to this procedure exhib-ited high efficiency and can be utilised for the separation oflow molecular weight molecules. The reproducibility studiesshowed that prepared columns can be used at high tempera-tures and pressures and are stable under these conditions withno evidence for monolith detachment from the titanium columnhousing.

The other important advantage of titanium housing is a betterheat dissipation not only during synthesis but also for chromato-graphic separation at elevated temperature, which is extremelyimportant for some separations where reasonable selectivitycan be achieved with applied column temperature gradients[44].

Acknowledgements

Authors would like to thank Science Foundation Ireland (GrantNumber 08/SRC/B1412) for research funding under the StrategicResearch Cluster programme, and Dr. Ken Cook and Dionex (U.K.)Ltd. for the provision of the UltiMate 3000 Rapid Separation LCsystem.

References

[1] F. Svec, A.A. Kurganov, J. Chromatogr. A 1184 (2008) 281.[2] F. Svec, T.B. Tennikova, Z. Deyl, Monolithic Materials: Preparation, Properties

and Applications, Elsevier, Amsterdam, 2003.[3] B. Paull, P.N. Nesterenko, Analyst 130 (2005) 134.[4] E.P. Nesterenko, P.N. Nesterenko, B. Paull, J. Chromatogr. A 1213 (2008) 62.[5] E.F. Hilder, F. Svec, J.M.J. Frechet, J. Chromatogr. A 1044 (2004) 3.[6] S. Xie, F. Svec, J.M.J. Frechet, Chem. Mater. 10 (1998) 4072.[7] C. Viklund, F. Svec, J.M.J. Frechet, K. Irgum, Chem. Mater. 8 (1996) 744.[8] A. Jungbauer, R. Hahn, J. Sep. Sci. 27 (2004) 767.[9] E. Gillespie, D. Connolly, B. Paull, Analyst 134 (2009) 1314.10] M.D. Goldberg, R.C. Lo, S. Abele, M. Macka, F.A. Gomez, Anal. Chem. 81 (2009)

5095.11] S.S. Liang, S.H. Chen, J. Chromatogr. A 1216 (2009) 2282.12] H. Minakuchi, K. Nakanishi, N. Soga, N. Ishizuka, N. Tanaka, Anal. Chem. 68

(1996) 3498.13] E.G. Vlakh, T.B. Tennikova, J. Sep. Sci. 30 (2007) 2801.14] Dionex Application Note 240, Dionex Corp, 2009.15] J. Courtois, M. Szumski, E. Bystrom, A. Iwasiewicz, A. Shchukarev, K. Irgum, J.

Sep. Sci. 29 (2006)14.16] E.C. Peters, M. Petro, F. Svec, J.M.J. Frechet, Anal. Chem. 69 (1997) 3646.17] P. Coufal, M.C. Cihak, J. Suchankova, E. Tesarova, Z. Bosakova, K. Stulic, J. Chro-

matogr. A 946 (2002) 99.18] L. Geiser, S. Eeltink, F. Svec, J.M.J. Frechet, J. Chromatogr. A 1140 (2007)

140.19] G.V. Lisichkin, Yu.A. Fadeev, A.A. Serdan, P.N. Nesterenko, P.G. Mingalyov,

D.B. Furman, Chemistry of Surface Grafted Compounds, Fizmatlit, Moscow,2003.

20] A. Nanci, J.D. Wuest, L. Peru, P. Brunet, V. Sharma, S. Zalzal, M.D. Mckee, J.Biomed. Mater. Res. 40 (1998) 324.

21] R. Padma, K. Ramkumar, M. Satyam, J. Mater. Sci. 23 (1988) 1591.22] D. Moravcova, P. Jandera, J. Urban, J. Planeta, J. Sep. Sci. 26 (2003) 1005.23] B.I. Ermolaev, Met. Sci. Heat Treat. 16 (1974) 1049.24] P. Bouchut, D. Decruppe, L. Delrive, J. Appl. Phys. 96 (2004) 3221.25] C.L. Choy, K.W. Kwok, W.P. Leung, F.P. Lau, J. Polym. Sci. B 32 (2003) 1389.26] E.C. Peters, F. Svec, J.M.J. Frechet, Chem. Mater. 9 (1997) 1898.27] F. Svec, J.M.J. Frechet, Macromolecules 28 (1995) 7580.28] V.V. Kireev, High-molecular Compounds, Vysshaya Shkola, Moscow, 1992.29] F. Svec, J.M.J. Frechet, Chem. Mater. 7 (1995) 707.30] K. Cabrera, J. Sep. Sci. 27 (2004) 843.31] D.T.T. Nguyen, D. Guillarme, S. Rudaz, J.L. Veuthey, J. Sep. Sci. 29 (2006)

1836.32] T. Teutenberg, J. Tuerk, M. Hozhauser, S. Giegold, J. Sep. Sci. 30 (2007)

1101.33] T. Andersen, Q.N.T. Nguyen, R. Trones, T. Greibrokk, Analyst 129 (2004)

191.

34] S. Eeltink, S. Dolman, F. Detobel, G. Desmet, R. Swart, M. Ursem, J. Sep. Sci. 32

(2009) 2504.35] T.J. Causon, R.A. Shellie, E.F. Hilder, Analyst 134 (2009) 440.36] Dionex Technical Note 82, Dionex Corp, 2009.37] E. Tyrrell, P.N. Nesterenko, B. Paull, J. Liq. Chromatogr. Relat. Technol. 29 (2006)

2201.

2 matog

[

[

[

146 E.P. Nesterenko et al. / J. Chro

38] C. Wohlfarth, Viscosity of Pure Organic Liquids and Binary Liquid Mixtures,Numerical Data and Functional Relationships in Science and Technology, vol.25, Springer, Berlin, 2009, pp. 97–99.

39] P. Holdsvendova, P. Coufal, J. Suchankova, E. Tesarova, Z. Bosakova, J. Sep. Sci.26 (2003) 1623.

40] A.M. Siouffi, J. Chromatogr. A 1126 (2006) 86.

[[[

[

r. A 1217 (2010) 2138–2146

41] L. Trojer, C.P. Bisjak, W. Wieder, G.K. Bonn, J. Chromatogr. A 1216 (2009) 6303.42] A. Greiderer, L. Trojer, C.W. Huck, G.K. Bonn, J. Chromatogr. A 1216 (2009) 7747.43] Y. Ueki, T. Umemura, Y. Iwashita, T. Odake, H. Haraguchi, K. Tsunoda, J. Chro-

matogr. A 1106 (2006) 106.44] L. Barron, P.N. Nesterenko, B. Paull, J. Chromatogr. A 1072 (2005) 207.